ES2375183T3 - AUDIO PLAYBACK, FEEDING SYSTEM AND METHOD. - Google Patents

Abstract

The audio reproduction apparatus comprises a cost input for inputting a mathematical cost derived from a measurement, which measurement is user-influenceable and a conditioning unit, capable of delivering an output audio signal in dependence of the mathematical cost, characterized in that the conditioning unit comprises an audio processing means arranged to process an input audio signal to derive the output audio signal with a reproduction quality in dependence of the mathematical cost. As a reproduction quality the position of a virtual sound source and the quality of a stereo signal are also possible. A system comprising the audio reproduction apparatus, a measurement device and a sound production device and a method of to deriving the output audio signal with a reproduction quality in dependence of the mathematical cost are also presented.

Description

Audio playback device, feedback system and method

The invention relates to an audio reproduction apparatus comprising:

-

 input means for entering an input audio signal;

-

 an output to output an output audio signal derived from the input audio signal;

-

 a cost entry to enter a mathematical cost derived from a measurement, a measurement that can be influenced by the user; Y

-

 a conditioning unit, which can deliver the output audio signal depending on the mathematical cost.

The invention also relates to an audio feedback system comprising:

-

 an audio source;

-

 a measuring device arranged to deliver a measurement that can be influenced by the user;

-

 a mathematical cost calculation unit, arranged to derive a mathematical cost from the measurement;

-

 a sound production device; Y

-

 a conditioning unit arranged to receive an input audio signal from the audio source, to receive the mathematical cost, and to deliver to the sound production device an output audio signal derived from the input audio signal, in dependence on the mathematical cost.

The invention also relates to a method for deriving an output audio signal from an input audio signal depending on a mathematical cost derived from a measurement that can be influenced by the user.

The invention also relates to a computer program for execution by a processor, which describes the method mentioned above.

The invention also relates to a data carrier that stores the computer program.

An embodiment of such an audio reproduction apparatus is known from US-A.

4,788,983. The Japanese patent summary for JP 10 207 340A discloses a system to generate a virtual space with audio and visual inputs. Document US 6 195 434 B1 discloses a system for generating a virtual sound source in a given position with respect to the user. US 4 110 918 A discloses a modular biofeedback training system. The apparatus known from US-A-4,788,983 is designed for use by a person performing a sporting activity, who wishes to listen to music. The known apparatus contains a conditioning unit that can transmit an input audio signal from a walkman as an output audio signal to headphones. The conditioning unit also receives a mathematical cost signal from a heart rate measuring device. A specific user, according to, for example, their age or sex, a safe heart rate window that they want to use during their training. If your heart rate is too low, you are obviously not getting enough exercise. On the other hand, if your heart rate is too high, your exercise could be harmful. The conditioning unit transmits the input audio signal only if the measured heart rate is within the desired window, and otherwise no sound is sent to the headphones.

It is a disadvantage of the known apparatus that such rudimentary regulation of the output audio signal is not easy for the user to understand. For example, if the window is narrow, it is difficult for a user to judge whether he has lost his music because he is running too slow or too fast.

An object of the invention is to provide an audio reproduction apparatus of the kind described in the initial paragraph, which is relatively versatile as regards the regulation of the output signal.

The object is realized because the conditioning unit comprises audio processing means arranged to process the input audio signal to derive the output audio signal with a reproduction quality depending on the mathematical cost. The conditioning unit of the known audio reproduction apparatus contains elements that only implement a switching function. In case the heart rate is outside the window, no signal is sent to the headphones. This is not very desirable.

If the user is training just a little soft, he will not hear absolutely any sound. Instead of motivating you to start running harder again, this can be demotivating for certain users. It may be desirable that there be a gradual change, so that the user can perform less than expected during a certain period and still have music. In addition, heart rate measurements are not always reliable, for example if a signal from a nearby second user is captured. The user is then punished for something beyond his control. The audio reproduction apparatus according to the present invention is arranged to offer the user many more versatile strategies to respond to his sporting activity, implemented as strategies to calculate the mathematical cost. The apparatus according to the invention is also arranged to provide more versatile output audio presentation strategies. Instead of simply interrupting the output audio signal, the apparatus of the invention offers options to change the playback quality of the output audio signal. It is a physical correlation that can be measured and determined with the psychoacoustic sound quality. For example, the audio reproduction apparatus can gradually reduce the amplitude of the sound, leading to less intelligible music. Alternatively, if stereo is available, a user may be punished who performs less than expected when delivering the audio reproduction device monophonic sound instead of stereo, in which case the number of independent output signals is a measure of reproduction quality of the perceived psychoacoustic quality.

In one embodiment, the reproduction quality comprises a three-dimensional position of a virtual sound source, the audio processing means being able to simulate the virtual sound source by means of the output audio signal. Of all the audio processing functions that can be applied to obtain a certain psychoacoustic quality, some perform the audio positioning of a virtual sound source in a three-dimensional space around a user's head. This is especially interesting for the user who travels a distance, real or virtual, such as a runner or a person riding an exercise bike. If you set the instruction on the device that you must run at a certain speed, it should be in a certain position, real or virtual, at a certain time. The audio reproduction apparatus can place the virtual sound source by sending an appropriate audio signal to a left or right speaker of the headphones, the virtual sound source being for example two virtual speakers at a distance of 1 meter in front of the user. If the user runs too slow, the virtual speakers will move away from him, and this can be simulated if desired by making the sound that emerges from the virtual speakers less audible. The user can reach the speakers by running faster. By adding synthetic reverbs, other three-dimensional audio position quality measurements can be influenced, such as the illusion of a wall in front of the user.

In a modification of the embodiment, the audio processing means comprise a filter arranged to simulate the virtual sound source position by deriving the output audio signal by filtering the input audio signal with a head-related transfer function ( HRTF) user dependent. Through a HRTF, a sound source such as a virtual speaker can be accurately located. The input audio signal for the right and left earphone loudspeaker respectively is filtered by a respective HRTF, simulating the path through a virtual space of sound from a real speaker in a position in the space to the user's respective ear.

In another modification of the embodiment or an additional variation in the previous modification, the audio processing means comprise an audio processing unit arranged to simulate the position of the virtual sound source by changing a property of the selected output audio signal. Signal amplitude and reverberation added. Both properties are simple signal processing functions to perform the illusion of sound with a specific three-dimensional audio position quality.

The audio processing means may also be arranged to derive a second output audio signal, together with the output audio signal constituting a stereo audio signal, the audio processing means being arranged to derive the stereo audio signal. from the input audio signal with a specified stereo quality dependent on the mathematical cost. The signal processing functions that influence stereo quality are, for example, the following:

-

 virtual speakers approach each other if the user performs less than expected;

-

 The audio signals for the left and right virtual speakers become more similar if the user performs less than expected; or

-

 One of the virtual speakers disappears if the user performs less than expected.

The last option can also be implemented as a switching, gradual or abrupt, between multichannel envelope such as Dolby 5.1 and 2-channel stereo. It is advantageous to vary the stereo quality given the extended presence of the stereo sound.

It is advantageous if the reproduction quality comprises a specification of the frequency distribution of the output audio signal. By changing the frequency content of the output audio signal, the audio processing means can simulate other effects. For example, a user is punished who performs less than expected by eliminating the bass from the audio signal. The three-dimensional position of a virtual sound source can also be influenced by such an audio processing function. For example, the audio processing means can be arranged to eliminate high frequencies as if the sound had to travel a long way through a dense fog, or was found at a depth in a virtual wall.

It is also advantageous if the audio reproduction apparatus comprises a first quality calculation unit for determining the reproduction quality for use in the subsequent derivation of the output audio signal by means of the audio processing means. Playback quality is a property of the output audio signal. The audio reproduction apparatus could measure it in the output audio signal, but then the input audio signal must first be processed by a processing algorithm unknown a priori. A correlation reproduction quality can be determined by the first quality calculation unit and sent to the audio processing means that apply a corresponding processing function. For example, changing the angle of two virtual speakers has an influence on stereo quality, and in general it is not necessary to specify what numerical stereo quality a user experiences. If more precision is desired, a specific angle function can be stored in a memory, for example based on user panel tests. It is also possible for the user to specify a relationship between, for example, his running speed and the angle himself, or choose between several pre-stored functions, some of which change the angle slowly and others quickly.

Alternatively or additionally it is advantageous if the audio reproduction apparatus comprises quality measurement means for measuring an output quality measurement of the output audio signal, and comprises means for calculating parameter value for calculating a parameter value, for use in the subsequent derivation of the output audio signal by means of audio processing means. If the quality of the output signal is measured, it can be feedback to change the processing in the input audio signal for future times. Such a feedback control loop obtains the desired reproduction quality after a certain tuning time. The parameter value calculation means can take into account an incompatibility between the result of the reproduction quality measurement and the desired quality measurement. The parameter value is changed accordingly, directing the processing function until the desired audio output signal playback quality is obtained.

In another embodiment, a mathematical cost calculation unit is included which is arranged to derive the mathematical cost from the measurement that can be received from a measuring device. A mathematical cost can be derived from any kind of apparatus, for example a random generator if the audio reproduction apparatus is used in a competitive game. Normally the mathematical cost can be determined based on a measurement that the user can influence, such as his running speed, heart rate, etc. A measuring device could send the mathematical cost directly to the device, for example as an encoded number. Normally, however, the audio reproduction apparatus may contain the new functionality, so that it can be used with a commercially available measuring device.

In a modification of the previous embodiment, the mathematical cost calculation unit is arranged to derive the mathematical cost based on a difference between the measurement and a chosen value. For example, the user sets this chosen value which is his desired running speed as 10 km / h, or his desired heart rate as 180 bpm. Then the quality is determined for example as the actual running speed minus 10 km / h. The faster you run, the more the playback quality changes. Or in an alternative version, if you run a little stronger nothing happens, and if you run much stronger the audio processing means are arranged so that the playback quality begins to change gradually depending on the amount of time you are running Stronger than the chosen value.

In an alternative modification of the previous embodiment or in addition to the previous modification, the mathematical cost calculation unit is arranged to derive the mathematical cost from a biometric measurement. Engineering quality audio systems and biometric measurements are unrelated fields of technology at all. Apparently, nobody thinks it necessary to combine them. Biometric measurement systems are usually designed by engineers working in close cooperation with doctors, and the priorities in this field are accuracy and robustness of measurements and safety. The quality of audio reproduction is more a matter of artistic taste. This has led to the fact that biometric measurements are usually displayed on numerical screens. An exception are the beeps of a medical monitor, but the audio functionality of such monitors is designed for simplicity rather than quality of artistic reproduction. There is a need for easy to understand feedback for the user of biometric data while training, since a user does not like to watch a screen continuously while training. However, when there is a sound present, it automatically enters the user's ear.

The audio feedback system is characterized in that the conditioning unit comprises audio processing means arranged to process the input audio signal for the derivation of the output audio signal with a reproduction quality depending on the mathematical cost. It is advantageous to produce the system as a whole, since then all components can be optimally tuned together.

The method of deriving an output audio signal from the input audio signal depending on the mathematical cost derived from a measurement that can be influenced by the user is characterized in that the output signal is derived with a reproduction quality specified dependent on the mathematical cost.

These and other aspects of the audio reproduction apparatus, the method, the audio feedback system, the computer program and the data carrier according to the invention will be apparent from and will be clarified with reference to the implementations and embodiments described below. in this document, and with reference to the attached drawings, which serve merely as non-limiting illustrations.

In the drawings:

Figure 1 schematically shows an application of the audio reproduction apparatus;

Figure 2 schematically shows an embodiment of the audio reproduction apparatus;

Figure 3 schematically shows an embodiment of the audio processing means of the audio reproduction apparatus;

Figure 4 schematically shows an embodiment of the audio feedback system;

Figure 5 schematically shows an embodiment of the data carrier;

Figures 6a and b schematically show a respective example of a mathematical cost function;

Figure 7 schematically shows an example specification of a position quality as a mathematical cost function; Y

Figure 8 schematically shows an example of a frequency spectrum of an output audio signal emitted by the audio reproduction apparatus.

In these figures the elements drawn in dashed lines are virtual in Figure 1, and optional in the other figures, depending on the desired embodiment. It is not necessary that all the elements present in the illustrative embodiment of the audio processing means in Figure 3 be present in an alternative embodiment.

Figure 1 shows a user 100 of the audio reproduction apparatus 200 according to the invention, who is jogging. For example, you could also be rowing on an indoor rowing machine. While practicing his sporting activity, he is listening to music that is proceeding, such as an output audio signal or, see Figure 2, of the audio reproduction apparatus 200, which is for example a portable audio player such as an MP3 player, and reproduced by the left and right earphone speaker 114 and 115 of a sound producing device 102, which are usually headphones. The playback quality R of the music is changed by the audio reproduction apparatus 200 depending on the sporting performance of the user 100, for example varying between Dolby 5.1 and mono without bass frequencies. For example, if you run too slow, you will be penalized with music of poor playback quality R. Your performance can be measured by at least one of several sensors. For example, a pedometer 108 connected to your sneaker or other measuring device such as a heart rate meter 106 delivers a measurement signal m. For the heart rate meter 106 this measurement can be for example a PQRST complex of an electrocardiogram, a temporal sequence of pulses, or a number representing the heart rate. The audio reproduction apparatus 200 converts this measurement m into a reproduction quality R.

In a simple variant of the audio reproduction apparatus 200, the audio reproduction strategy is set, and the user 100 can only specify the way in which the measurement is transformed into a mathematical cost c. Note that for reasons of clarity the examples are described for an embodiment in which all mathematical transformations are performed as software algorithms that are executed in a processor, but a dedicated hardware circuitry could also be used. For example, user 100 may specify an iv interval in which his heart rate must remain as [LL, LU] in Figure 6a. A mathematical cost function 602 is shown in a coordinate system 600 with on the x-axis the heart rate measurement m minus a chosen value d and on the axis and the mathematical cost c. This chosen value d is set by the user as the target heart rate for training, for example 180 bpm. The interval [LL, LU] can be symmetric or asymmetric around the chosen value d. The mathematical cost function can be set in the hardware of the audio reproduction apparatus 200 of Figure 2, or the user can specify by means of user interface means 311 of Figure 3 how the mathematical cost changes with m-d. For example, as in Figure 6a until the marker values ML and MU the mathematical cost changes linearly with a small inclination, while between the marker values and the limits of the interval iv the mathematical cost is inclined with greater slope, and outside the interval iv the mathematical cost increases with a lot of slope. In a specification 606 of the alternative mathematical cost function shown in Figure 6b, the mathematical cost is non-zero and changes linearly only outside the training interval iv. For example, in Figure 6b, user 100 has designed a cost function with negative values if it is running too slow. Then a negative cost can be easily mapped during playback, for example as a negative angle to a virtual sound source, and a positive cost corresponding to speeds that are too high as a positive angle a. In this way, both cases can be easily discriminated. The user is free to design the cost function by choosing, for example, training interval limits, speed of cost change, which can result in speed of change of audio reproduction quality R, if only the measurements below the value chosen d lead to a non-zero cost, etc.

The user interface means 311 is for example software that runs on a processor, which asks the user to type numerical values of marker values and inclinations, or that allows the user to draw the mathematical cost function 602 graphically. The audio reproduction strategy set is for example the one illustrated in Figure 7. In this case the reproduction quality R is implemented as what is called in the rest of the text a reproduction quality of position P, which is any specification of a physical parameter of the output audio signal or that results in the perception that the output audio signal comes from a position of a virtual sound source. For example, the virtual sound source can be perceived near a user's head or away, or in Figure 7 it is an angle a of a virtual sound source around a user's head. If the user 100 is running with the desired target speed, the mathematical cost c is zero and the angle a is also zero degrees, ie the virtual sound source is directly in front of the user 100. If the runner runs too slow or fast , the mathematical cost c with a specification like that of Figure 6b decreases or increases respectively, and the virtual sound source changes to the left and right side respectively of the user 100. The sound source can remain behind the user until the user 100 run with the desired speed of the chosen value d again or run with the desired speed again for a certain amount of time. Alternatively, the sound source can also start behind user 100, being an annoying motivation to run harder.

In more advanced variants of the audio reproduction apparatus 200, the user 100 can also specify the strategy to change the reproduction quality as a function of the mathematical cost c. You can program a first quality calculation unit 330 in Fig. 3 which emits for example as stereo quality S an angle 160 between a first virtual sound source 152 and a second virtual sound source 154 generated by the output audio signal oy a second output audio signal o2, as a linear function of the mathematical cost c. Or you can select an alternative or additional audio processing function, which for example adds a reverberation amount as a function of the mathematical cost c to simulate a virtual speaker distance in a virtual space.

Another example application of the audio reproduction apparatus 200 is the prevention of repetitive strain injury (RSI) or inactivity. In this case the user 100 sits for example in front of a personal computer (PC) or on a sofa in front of a television (TV). The sound producing device 102 is for example a speaker connected to the PC or television. The chosen value d is the amount of time the user 100 wishes to work or watch TV before resting. The mathematical cost c is determined for example by the amount of time t elapsed since a start time t0 minus the chosen value d, the permissible time to work or watch TV continuously:

c = (t-t0) -d if t-t0> d;

c = 0 if t-t0 <d [1]

With the speakers of a TV or a PC, a virtual sound source position can be simulated.

Figure 2 schematically shows an embodiment of the audio reproduction apparatus 200 in its basic form. An input audio signal i enters through input means 204 from for example a portable MP3 player or the sound card of a PC. The input audio signal i can proceed from outside or inside the audio reproduction apparatus 200. In the latter case the audio reproduction apparatus 200 may comprise for example a CD player unit or any internal audio source 201. The input audio signal i can be mono or multichannel audio. There is also a cost entry 208 for receiving a mathematical cost c from a mathematical cost calculation unit 210, which is arranged to derive the mathematical cost from a measurement performed by a measuring device 212. The mathematical cost calculation unit 210 may be incorporated in the audio reproduction apparatus 200 or may be separated, for example in the measuring device 212. Normally, the measuring device 212 is not incorporated in the audio reproduction apparatus 200, although it could be, in case it is for example a clock. The audio reproduction apparatus 200 contains a conditioning unit 202, which contains audio processing means 216 arranged to process the input audio signal i to derive the output audio signal or with a dependent reproduction quality R of the mathematical cost c. The output audio signal or goes to an output 206, to which a speaker 214 can be connected. The audio processing means 216 can simply perform a single parametric function leading to an output audio signal or playback quality signal. variable and therefore of perceptible psychoacoustic quality, or multiple audio processing algorithms can be applied alternatively or simultaneously as in Figure 3.

Figure 3 schematically shows audio processing means 316, which is an embodiment of the audio processing means 216 of the audio reproduction apparatus 200. In the audio processing means 316 several processing units are shown merely to explain various characteristics of the audio reproduction apparatus 200, and it should be apparent that other combinations are possible. The audio processing means 316 is arranged to provide an output audio signal

or to a first speaker 314 and if a second output audio signal o2 to a second speaker 315 is required.

For many audio processing algorithms, the reproduction quality R can be set in advance, and subsequent audio processing is chosen depending on the reproduction quality R. For example the reproduction quality R can be a parameter of a processing algorithm of audio, as in the case where the amplitude of the output audio signal is established. This can be done with a variable gain amplifier. In other cases, user panel tests or the preferences of the actual user of the audio reproduction apparatus 200 can be used to select an appropriate processing algorithm, for example the first, second or third processing algorithm 320, 322 or 324, respectively . In the exemplary embodiment of Figure 3, a third mathematical cost c3 from a second measuring device 352 goes to an algorithm selector 326, which for example contains an interval table. If the third mathematical cost c3 falls within a first interval, the first processing algorithm 320, etc. is selected. Such a configuration makes it possible to switch to totally different algorithms depending on the value of the third mathematical cost c3. For example, the first algorithm can change the angle 160 between two virtual speakers, depending on where the third mathematical cost c3 is located in the first interval. If the third mathematical cost c3 becomes so high that it is outside the first interval and in a second interval, the second processing algorithm 322 is selected, which for example changes the amplitude of the signals from the virtual speakers, or both the angle 160 between them as the amplitude of the signals. Another example in which the reproduction quality R is set before processing is the configuration of an angular position in a sphere around the head of the user 100 of a virtual sound source by means of a transfer function related to the HRTF head . For example, when the user 100 uses headphones, the input audio signal i can be simulated to come from a virtual sound source position, filtering it by means of filter 332 using a first specific HRTF to obtain the output audio signal or for the left earphone speaker 114 and using a second specific HRTF to obtain the second output audio signal o2 for the right earphone speaker 115. Both HRTFs are dependent on the required position of the virtual sound source, for example specified as two angles in a unit sphere, and can be taken from a memory 334 containing HRTF for several different positions. A first quality calculation unit 330 determines the reproduction quality R necessary for additional audio processing. For example, in the case described above, the first quality calculation unit 330 calculates an angle a of the virtual sound source, used to take the HRTFs, as a linear function of a first mathematical cost c1. The first mathematical cost c1 is derived from a measuring device 312 by a mathematical cost calculation unit 310, which for example evaluates a function such as that of Figure 6a. Details on HRTF measurement can be found in WO 01/49066 and the document "F.L. Wightman and D.J. Kistler: Headphone simulation of free field listening. I: Stimulus synthesis. Journal of the Acoustical Soc. Of America 85 # 2, February 1989, pages 858-867 ”.

In other cases, the reproduction quality R in the output audio signal or in itself has to be measured, for example because the relationship between the reproduction quality R of the output audio signal or and the particular processing is too complex of formulate or be unknown. In this case the feedback can be used to select the correct processing algorithm or the correct parameter for a parametric processing algorithm. The quality measurement means 344 measures an output quality measure M of the output audio signal or. The output quality measure M and a desired reproduction quality R from a second quality calculation unit 340 are fed to means 346 for parameter value calculation. From these two parameters, a pv parameter value is calculated to direct the subsequent processing by an audio processing unit, which selects a particular processing algorithm or changes a parameter of a parametric algorithm. This can be done by any technique known to control theory. For example, an update parameter value pv can be calculated as in equation [2]:

pv = 8 (M-R) [2]

Actually the parameter value pv can be any function of M and R, if necessary also taking into account that the measurement of output quality M is a function different from the desired psychoacoustic quality than the reproduction quality R.

The implicit functionality required between the output quality measure M and the reproduction quality R derived from the first mathematical cost c1, can be specified by the user 100. With the user input means 360, for example a keyboard, a panel sensitive to the touch, or a rotary knob and user interface means 311, user 100 can specify several desired measurement values d, which become corresponding first mathematical costs c1 and reproduction qualities R. Instead of entering desired measurement values d , user 100 can also enter desired mathematical costs cs. During this learning phase, for each reproduction quality R several processing algorithms are explored with corresponding output quality measures M. When the output quality measure M corresponding to a psychoacoustic quality is reached as desired by the user 100 , the user 100 can indicate this to the means 346 of parameter value calculation through a learning control connection lc, wired or wireless. Then, the parameter value calculation means 346 can store the reproduction quality pair R and parameter value pv, so that during operation, feedback is no longer needed, although instead, from a quality R corresponding to a measurement m, the correct parameter value pv can be sent to an audio processing unit 342. The lc learning control connection can also be used to specify what available processing should be used to obtain an output audio signal

or with the desired reproduction quality R, setting an output selector 370.

Another example of setting user preferences in audio processing means 316 is illustrated with a second user input means 361. In this example, user 100 can directly enter a second desired mathematical cost cs ’and a corresponding desired processing algorithm selection na into the algorithm selector 326.

Note that in many cases the exact perceived psychoacoustic quality, for example the exact position of a virtual sound source is not important, but only that the psychoacoustic quality changes monotonously. This mitigates the requirements in the reproduction strategy. Any virtual sound source angle mapping for the mathematical cost for example might already be enough.

Many processing algorithms can be designed to create a certain perceptible psychoacoustic quality that corresponds to a reproduction quality R that characterizes a selected algorithm. For example, a bass reproduction can be changed depending on the mathematical cost. As shown in Figure 8, as a particular reproduction quality R or part of a reproduction quality R, a SPEC specification that reflects the spectral content of the output audio signal or can be calculated. An example of the SPEC specification is a frequency below which substantially no sound energy is present, for example a first low frequency FL1 or a second low frequency FL2. For example, if the user runs at almost the desired speed, bass is played until the first low frequency FL1. However, if you run too slow, you lose the bass between the first low frequency FL1 and the second low frequency FL2. Another example of the SPEC specification is the percentage of energy in the bass range [FL1, FL2] compared to the energy in the range [FL3, FH]. Any equalization strategy can be used as a function of cost c, for example the amount of treble can be a function of cost c.

An interesting algorithm establishes the mathematical cost function by means of a three-dimensional bubble 150 around the user's head 100. If it runs too slow, the distance between the virtual position of its head 158 and a mark point 156 in the bubble increases, which leads to an increased mathematical cost c, and a decreased reproduction quality R. The reproduction quality R can also change depending on whether the user 100 is within the bubble 150 or not, which facilitates a training tolerance . The virtual movement of the bubble 150 compared to the running action of the user could even keep track of whether the user waits at a traffic light, identifying this situation for example when he presses a button. When inside the bubble 150, the sound can sound for example as if the user 100 were in a particular space, selecting the HRTF corresponding to that space, while outside the bubble 150, the sound sounds dull.

The distance of a virtual sound source can also be simulated. The audio processing unit 342 for example can simulate the distance of the virtual sound source by changing the amplitude of the output audio signal or and the second output audio signal o2. Or reverberation can be added. In a space if a sound source is close, there is little reverberation, while if the sound source is far away, there is a greater proportion of reverberation. Additionally or alternatively, virtual corridors can be generated, in which the virtual sound source reflects its sound.

Several options are also possible to change an S stereo quality of a stereo audio signal. For example, an angle 160 between a first virtual sound source 152 and a second virtual sound source 154 may decrease as the mathematical cost c increases, resulting in a lower stereo quality S. Or the first and second audio signal of output o and o2 respectively can be made more similar. Or there may be a rigid switching from stereo to mono. A stereo signal must also be understood as multichannel audio, and a corresponding stereo quality S is for example the number of channels.

An example of a gradual change between stereo and mono is done by calculating changed stereo signals L ’and R’ according to the following equations:

LR

L '= (1+ a) + (1-a)

LR

R '= (1-a) - (1+ a) [3]

By changing the value of a parameter to between 0 and 1, a change between mono and stereo is made.

The measuring device 312 and the second measuring device 352 can be anything, for example a watch or a GPS sensor indicating the position of the user 100. In particular it could be a biometric measuring device, such as a pedometer connected to a shoe, a chest strap or heart rate meter 106 for the earlobe, a thermometer, etc. These measuring devices are interesting when used for measuring sports performance. The measuring device 312 could also be incorporated into a professional training apparatus, for example an endless belt for jogging indoors. Running style, particularly if there are health problems, could also be fed back.

In a game-like application, a second mathematical cost c2 can be established by means 313 that determine the cost, for example a random generator. Depending on its fate, the mathematical cost of the user 100 is set back to the second mathematical cost c2 and has to run harder to regain the level of the third mathematical cost c3, as determined by a second calculation unit 350 of mathematical cost from the second measuring device 352.

Figure 4 schematically shows an embodiment of the audio feedback system. An audio source 421 delivers an input audio signal i to a conditioning unit 402. The audio source 421 can be for example a portable audio player or an audio distribution server in a gym. A measuring device 412 performs a measurement m, which is converted by a mathematical cost calculation unit 410 into a mathematical cost c, which is also introduced into the conditioning unit 402. The conditioning unit 402 contains audio processing means 416, which are arranged to process the input audio signal to obtain an output signal or of a reproduction quality R dependent on the mathematical cost c, which is sent to a means 414 sound production, for example a pair of headphones or the speakers of a television. The measuring device 412, the mathematical cost calculation unit 410, the conditioning unit 402, the audio source 421, and the sound production means 414 can be performed separately or in any combination. For example, normally an audio reproduction apparatus may contain the conditioning unit 402, the audio source 421 and the mathematical cost calculation unit 410.

A head tracking device may be present, so that the positions of the virtual sound sources for movements of the user's head 100 are corrected. Means, for example a microphone, may also be present to capture certain sounds from the environment. and mix them with the signal for the sound production device 414 to improve safety.

The specification of, for example, the cost function instead of being performed by the user 100 may come from a channel, such as the Internet, through an interface, from a second person, for example a personal trainer. Or the specification can be made at a different time by user 100, for example on his PC. For example, you can perform a training schedule for the whole month, which can be downloaded, for example, wirelessly to the audio playback device. Function parameters and functions can even be downloaded from, for example, the Internet, and for example shared between users who want similar training sessions. Specifications made during training and stored in a memory can also be downloaded through the interface to the computer for further analysis, for example training enhancement.

It should be noted that the above-mentioned embodiments illustrate instead of limiting the invention and that those skilled in the art can design alternatives, without departing from the scope of the claims. Apart from the combinations of elements of the invention as combined in the claims, the invention covers other combinations of the elements within the scope of the invention as perceived by one skilled in the art. Any combination of elements can be done in a single dedicated element. It is not intended that any reference sign in parentheses in the claim limit the claim. The term "comprising" does not exclude the presence of elements or aspects not listed in a claim. The word "a"

or "one" before an element does not exclude the presence of a plurality of such elements.

The invention can be implemented by means of hardware or by means of software executed on a computer, and previously stored on a data carrier or transmitted by a signal transmission system.

Claims (13)

1. Audio playback apparatus (200) comprising:

-

 input means (204) for entering an input audio signal (i);

-

 an output (206) to output an output audio signal (or) derived from the input audio signal (i);

-

 a cost entry (208) for entering a mathematical cost (c) derived from a measurement (m), measurement (m) that can be influenced by the user; Y

-

 a conditioning unit (202), which can deliver the output audio signal (o) depending on the mathematical cost (c),

characterized in that the conditioning unit (202) comprises audio processing means (216, 316) arranged to process the input audio signal (i) to derive the output audio signal (o) with a reproduction quality ( R) depending on the mathematical cost (c); and characterized in that the reproduction quality comprises at least one of:

(to)

 a specified stereo quality of the output audio signal that is a stereo audio signal;

(b)

 several independent output signals;

(C)

 a specification of a frequency distribution of the output audio signal; Y

(d)

 an amplitude of the output audio signal.

2.

Audio reproduction apparatus (200) according to claim 1, wherein the reproduction quality (R) comprises a three-dimensional position of a virtual sound source (152), the simulation audio processing means (216, 316) being able to the virtual sound source (152) by means of the output audio signal (o).

3.

Audio reproduction apparatus (200) according to claim 2, wherein the audio processing means (316) comprise a filter (332) arranged to simulate the position of the virtual sound source (152) by deriving the audio signal output (or) by filtering the input audio signal (i) with a user-dependent head-related transfer function (HRTF).

Four.

Audio reproduction apparatus (200) according to claim 2, wherein the audio processing means (316) comprises an audio processing unit (342) arranged to simulate the position of the virtual sound source (152) by changing a property of the selected output audio signal (or) of added signal amplitude and reverberation.

5.

Audio reproduction apparatus (200) according to claim 1, comprising a first quality calculation unit (330) for determining the reproduction quality (R) for use in the subsequent derivation of the output audio signal (or ) by means (216) of audio processing.

6.

Audio reproduction apparatus (200) according to claim 1, comprising quality measurement means (344) for measuring an output quality measure (M) of the output audio signal (o), and comprising means ( 346) parameter value calculation to calculate a parameter value (pv), for use in the subsequent derivation of the output audio signal (or) by means of audio processing means (216).

7.

Audio reproduction apparatus (200) according to claim 1, wherein a unit is comprised

(310) of mathematical cost calculation that is arranged to derive the mathematical cost (c) from the measurement (m) that can be received from a measuring device (312).

8.

Audio reproduction apparatus (200) according to claim 7, wherein the mathematical cost calculation unit (310) is arranged to derive the mathematical cost (c) based on a difference between the measurement (m) and a chosen value (d).

9.

Audio reproduction apparatus (200) according to claim 7, wherein the mathematical cost calculation unit (310) is arranged to derive the mathematical cost (c) from a biometric measurement.

10.

Audio feedback system comprising:

-

 an audio source (421); - a measuring device (412) arranged to deliver a measurement (m) that can be influenced by the user;

-

 a mathematical cost calculation unit (410), arranged to derive a mathematical cost (c) from the measurement (m);

-

 a sound production device (414); Y

-

 a conditioning unit (402) arranged to receive an input audio signal (i) from the audio source (421), to receive the mathematical cost (c), and to deliver to the sound producing device (414) a output audio signal (o) derived from the input audio signal (i), depending on the mathematical cost (c),

characterized in that the conditioning unit (402) comprises audio processing means (416) arranged to process the input audio signal (i) to derive the output audio signal (o) with a reproduction quality (R) depending on the mathematical cost (c); and characterized in that the reproduction quality comprises at least one of:

(to)

 a specified stereo quality of the output audio signal that is a stereo audio signal;

(b)

 several independent output signals;

(C)

 a specification of a frequency distribution of the output audio signal;

(d)

 an amplitude of the output audio signal.

11. Method to derive an output audio signal (o) from an input audio signal (i) depending on a mathematical cost (c) derived from a measurement (m) that may be influenced by the user, characterized in that the output signal (o) is derived with a specified reproduction quality

(R)

 dependent on the mathematical cost (c); and characterized in that the reproduction quality comprises at least one of:

(to)

 a specified stereo quality of the output audio signal that is a stereo audio signal;

(b)

 several independent output signals;

(C)

 a specification of a frequency distribution of the output audio signal;

(d)

 an amplitude of the output audio signal.

12.

Computer program for execution by a processor, which describes the method of claim 11.

13.

Support (500) of data stored in the computer program of claim 12.